My prediction on 3.6

[J Grant]

Desperate appraisers keep falling into the same trap. Do X, because if you don't, the client will find another appraiser to do it!!. Sounds like the defense drug dealers make - if I didn't sell them dope, somebody else would sell it.

How did the compromising and taking cheap fees and or hitting value to keep a client work out ? Not well, did it.

Using AI without verifying the results and without searching for our own comps and testing the adjustments would be evidence that the appraisers are lazy, and thus, not necessary. Perhaps that is why the GSE's are not so subtly encouraging appraisers to try all software options - appraisers following a predictable fear-monger pressure of let AI do the report fast or you can't; compete will be easily used against them. Appraisers have been marginalized; the next step is to eliminate them. If an appraiser lets the AI do it all, clearly the appraiser can be replaced with a "properly trained" analyst or RE agent who can guide AI a bit in filling out a 3.6 UAD- just as they are letting RE agents take a quick training course to become a "professionally trained data collector" and replace the appraiser on the inspection. A few appraisers on the review or management end would be all they need to call it "appraiser-verified" or some such label.

If it goes well and appraisers use the available AI responsibly, maybe our role will continue. If it does not go well, and the last segment of lender work is gone, maybe an appraisal as a viable profession will mean a commercial license going forward, with very few real estate licenses. A cert gen can do either type of property so it makes sense.

 

[HVNY]

After 25 years appraising, listing, selling real estate in a 4 county radius I AM AI. Also google maps. I can tell you where the good coffee is, the best egg sandwiches, hiking trails, parks, etc. I'm not going to babysit AI.

 

[J Grant]

AI only has access to the same data we have , the same sales and listings we have access to. All it does is process the data ultra-fast. AI suffers no consequences for what it spits out. Its career and license are not on the line. It does not suffer the loss of overpaying or overborrowing. It is a computer-driven software, disguised as having "intelligence."

That is why twenty-year-old tech guys who know zero about real estate are bringing out products touted to "revolutionize appraisals." They don't; even know what an appraisal is,-it;s just a money grab. It must be simple on their end to design an AI-based program to interact with MLS or public record data - such programs already exist, such as Spark or Synapse - all they have to do is tweak it. Ironically, these self-proclaimed genius guys failed to do their research to find that the number of real estate appraisers is declining in lender work, with many being too broke to buy their product.

In which case, they will pedal their AI appraisal programs to RE agents. I can just imagine the AI-enhanced CMAs the agents will hand to buyers. (not much different than the cherry picked and weird sales they provide now). Maybe the AI will add a narrative to a CMA to convince buyers to overpay.

These start-ups need to expand their customer base, or they will exit as quickly as they arrived.

 

[HVNY]

I'm working on a 4 plex that is defined on the public record as a 2 family. Nuff said...

 

[nachocheesefries]

Thanks- I believe I have been inputting my viewpoint throughout the thread - that appraisers need to maintain standards whether they rely on AI as an assist or not. It sounds too tempting to simply let AI do it, and if that is the case then truly the appraiser is no longer in charge of their own work product.

I don't think market forces are going to make it possible to exist in the residential space without using AI...so, yes, we are gong to have to learn how to work with AI and be compliant. I worry that the early adopters are gonna run away with it and leave the most experienced appraisers behind.

 

[nachocheesefries]

Desperate appraisers keep falling into the same trap. Do X, because if you don't, the client will find another appraiser to do it!!. Sounds like the defense drug dealers make - if I didn't sell them dope, somebody else would sell it.

How did the compromising and taking cheap fees and or hitting value to keep a client work out ? Not well, did it.

Using AI without verifying the results and without searching for our own comps and testing the adjustments would be evidence that the appraisers are lazy, and thus, not necessary. Perhaps that is why the GSE's are not so subtly encouraging appraisers to try all software options - appraisers following a predictable fear-monger pressure of let AI do the report fast or you can't; compete will be easily used against them. Appraisers have been marginalized; the next step is to eliminate them. If an appraiser lets the AI do it all, clearly the appraiser can be replaced with a "properly trained" analyst or RE agent who can guide AI a bit in filling out a 3.6 UAD- just as they are letting RE agents take a quick training course to become a "professionally trained data collector" and replace the appraiser on the inspection. A few appraisers on the review or management end would be all they need to call it "appraiser-verified" or some such label.

If it goes well and appraisers use the available AI responsibly, maybe our role will continue. If it does not go well, and the last segment of lender work is gone, maybe an appraisal as a viable profession will mean a commercial license going forward, with very few real estate licenses. A cert gen can do either type of property so it makes sense.

You must not be getting inundated with the same emails I am, asking 'click here' if you are ready for 3.6 and want to be added to the list. They are actively replacing slow adopters and relegating them to the portfolio panel...which is fine except there is not enough work for everyone on the portfolio side.

I would argue the alamode typewriter side are the ones fear mongering over compliance.

 

[chad hampton]

You are correct, sir!

That is crazy that leadership in this profession came out and said experience doesn’t really matter. I can’t imagine another profession saying something like that.

 

[chad hampton]

This is what you get with the corporate mentality. They try to put on a professional front. they hire some people that make promises and talk good on camera and all that. And they usually hire somebody with credentials to help them look legitimate. But when you get down to the nuts and bolts of it, they are high volume production outfits that are into churning and burning through appraisals. And they also burn through their appraisers pretty quick. It’s been that way with appraisal Mills and AMCs for as long as I can remember.

 

[George Hatch]

I came up dictating narratives (learned how to dictate back when I was writing police reports).

If I had all my research and analysis done before I started my report I could dictate a fairly detailed narrative report in about an hour or so. After that, I could proofread, edit and do any tweaking on the analysis in less than 1/2 hour and sometimes less than that. And those reports are way more SR1/SR2 complicated than SFR reports due to developing multiple approaches to value and considering other elements.

The reading and checking the analysis part of reviewing goes WAY faster than writing the report and then proofing. And, I could more readily catch things I missed. Since I didn't spend hours physically writing the report I wasn't anticipating so much what I would be seeing. New-to-me, so to speak.

The reason I don't dictate reports any more is because I'm not doing the same 1 or 2 different property types over and over again and it's harder to find a groove when you're only doing a certain property type every couple of months or so. And because it's not easy to get all the analysis together and in order prior to writing the report. Not to mention the point that the reason we use boilerplate is for elements of the report that usually don't vary from one report to another. Again, faster and easier to edit a boilerplate than to write it from scratch.

Some years ago the DataMaster users told me that app was saving them an hour in data entry, even though they sometimes had to tweak some of the details or fill in for omissions.

When I review and am verifying comparable data that process only takes a few minutes to call each of the transactions up and check the details.

My point is that if an AI is populating a boilerplate and even if its doing some of the calculations and summaries, its going to be pretty easy for the appraiser to come back in and clean things up or add more content and commentary. In the event there are errors and omissions, which may not always be the case.

Look at how many of us are quoting AI summaries in lieu of doing our own thinking and writing, or as a starting point for adding our own opinions. That's not a criticism. Heck, AI summaries show up in a lot of search engine results, some being more on point than others. IMO you can't say that you think the machine is unreliable when you're already relying on it on a regular basis.

 

[Sadie]

IDK- I took the first day of an online training webinar with ALAMODE today....boring as heck - it is doable but a learning curve. I plan on, after part two tommorow, doing a UAD 3.6 on a house I already appraised in 2.6 and see how it goes.

There is a way to start a report in 2.6 and then merge it with 3.6, whether that saves time or preserves things idk. There is a way to make a template which he has not covered yet and a way to save comments so it is not hopeless. Just a lot more detail and fields to fill out.

I have been beta resting ACI and found the easiest thing to do was write it up either in a word document or in the 2.6 and copy/paste it over. I had been doing the grid in the 2.6 also because you can't see how the adjustments affect the comparables in real time. Things may have changed since I last tested it though. I was pretty much writing 2 reports for each property.